Keap1-Nrf2 Pathway Regulates ALDH and Contributes to Radioresistance in Breast Cancer Stem Cells

Tumor recurrence after radiotherapy due to the presence of breast cancer stem cells (BCSCs) is a clinical challenge, and the mechanism remains unclear. Low levels of ROS and enhanced antioxidant defenses are shown to contribute to increasing radioresistance. However, the role of Nrf2-Keap1-Bach1 signaling in the radioresistance of BCSCs remains elusive. Fractionated radiation increased the percentage of the ALDH-expressing subpopulation and their sphere formation ability, promoted mesenchymal-to-epithelial transition and enhanced radioresistance in BCSCs. Radiation activated Nrf2 via Keap1 silencing and enhanced the tumor-initiating capability of BCSCs. Furthermore, knockdown of Nrf2 suppressed ALDH+ population and stem cell markers, reduced radioresistance by decreasing clonogenicity and blocked the tumorigenic ability in immunocompromised mice. An underlying mechanism of Keap1 silencing could be via miR200a, as we observed a significant increase in its expression, and the promoter methylation of Keap1 or GSK-3β did not change. Our data demonstrate that ALDH+ BCSC population contributes to breast tumor radioresistance via the Nrf2-Keap1 pathway, and targeting this cell population with miR200a could be beneficial but warrants detailed studies. Our results support the notion that Nrf2-Keap1 signaling controls mesenchymal–epithelial plasticity, regulates tumor-initiating ability and promotes the radioresistance of BCSCs.


Introduction
Radiotherapy (RT) is a critical factor of primary, adjuvant and palliative treatment for almost all kinds of cancers, including breast cancer. It alone is capable of lowering the 10-year risk of relapse by one half and reducing the 15-year risk of breast-cancer-related death [1]. Although profound benefits are achieved with RT due to its localized treatment, especially for ductal carcinoma and early invasive cancer, local control of the disease fails by 8-15% in radiotherapy-treated patients with advanced invasive tumors due to resistance and relapse of the tumor [2]. The reason for RT failure and the locoregional recurrence of breast cancer is the presence of a subset of radioresistant tumor cells, termed breast cancer stem cells (BCSCs), which show a difference in sensitivities to radiation [3][4][5]. Standard fractionated doses of radiation are sublethal for BCSCs as they typically evade radiation to develop innate or acquired resistance and establish tumor recurrence and metastasis, leading to the majority of cancer-related deaths. The molecular mechanisms that govern the emergence of aggressive radioresistance in BCSCs are yet unknown.
Low levels of reactive oxygen species (ROS) and enhanced ROS defenses appear to partially contribute to the adaptive tumor radioresistance in BCSCs [5][6][7]. Thus, the identification of underlying mechanisms and overcoming low ROS levels within BCSCs may be a useful method for improving radiation therapy. The transcription factor nuclear factor (erythroid-derived 2)-like 2 (Nrf2), the master regulator of antioxidant defense mechanisms, is a critical regulator of the redox balance. In the cytosol, Nrf2 activity is tightly regulated by two main inhibitors, Keap1 and GSK-3. The Neh2 and Neh6 domains of Nrf2 are the degron.

Mammosphere Formation
After irradiation, mammospheres were formed using single-cell suspensions in an ultralow attachment 6-well plate at a density of 2 × 10 4 cells/well in specific media for mammosphere culture containing DMEM and Nutrient Mixture F-12 medium supplemented with 20 ng/mL EGF, B27 (1:50, Life Technologies, MA, USA), 20 ng/mL FGF (Sigma-Aldrich, St. Louis, MO, USA) and penicillin/streptomycin (Invitrogen) for 4-5 days. The floating aggregates with a >50 µm diameter were selected as mammospheres, manually counted and dissociated by incubation with 1:5 of 0.25% trypsin/EDTA. The mammosphere-forming efficiency (MFE) was calculated using the following equation: MFE = No. of spheres formed/No. of cells seeded × platting efficiency.

Colony Formation
For the colony formation assay, MCF-7 cells (1000 cells/well) were grown on 6-well plates and maintained in a humidified chamber comprising 95% air and 5% CO 2 at 37 • C for 14 days. Cells were then fixed with 3.7% paraformaldehyde at room temperature for 10 min and stained using crystal violet solution (0.2% crystal violet and 1X PBS) at room temperature for 30 min. Stained cells were washed with 1× PBS and air-dried at room temperature. The numbers of colonies were quantified using the Image J program (version-v1.53e). The survival fraction was calculated as the number of colonies counted/the number of cells inoculated × plating efficiency at 0 Gy. Colonies consisting of 50 or more cells were counted as clonogenic survivors.

FACS Analysis for CD44/24 and ALDEFLUOR Assay
Cells were irradiated with the specific doses of radiation and then stained with anti-CD44-APC and anti-CD24-PE with their respective isotype controls, incubated for 40 min and analyzed on a FACSCanto™ flow cytometer (BD). To measure ALDH activity, cells were analyzed by an ALDEFLUOR assay kit (STEMCELL Technologies), following the manufacturer's protocol. Flow cytometry was performed on a FACSCanto™ flow cytometer (BD), USA and analyzed by DIVA software (BD Biosciences, USA).

qRT-PCR
Total RNA was extracted using TRIzol (Invitrogen). cDNA was synthesized using reverse transcription, followed by quantitative real-time PCR with SYBR Green Supermix (Life Technologies), using primers for Nrf2, Keap1, HO1, NQO1, SOX2, NANOG, KLF4 and GAPDH, which were used as the normalizing control. miR200a detection was carried out using the stem-loop method, as described previously [16]. The gene-specific primers used to perform real-time qRT-PCR analysis are listed in Table S1.

Nrf2 Activity
Nuclear protein fractions of irradiated MCF-7 cells were isolated using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher, MA, USA), according to the manufacturer's instructions. Protein concentrations were assessed using the Bradford reagent (Bio-Rad). Nrf2 Transcription Factor Assay (Colorimetric) was performed using 20 mg of nuclear proteins (ab207223, Abcam) to detect nuclear Nrf2 and antioxidant responsive element (ARE) sequence binding at OD 450 nm, following the manufacturer's instructions.

Scratch Wound Assay
MCF-7 cells were irradiated with specific doses of radiation and incubated for 24 h. The cells were scratched with a pipette tip to create wounds. Images were taken at different planes at 0 h and 24 h at 10× magnification. Percent cell migration was calculated as described in our previous paper [17].

ROS Detection
Detection of ROS was performed, as described previously [18]. Cells were treated with 1 µmol/L 2 ,7 dichlorodihydrofluorescein diacetate (DCF-DA; Invitrogen, MA, USA) for 30 min, followed by a 1× PBS wash for 2 times. The reduced DCF-DA was oxidized by intracellular ROS and converted into fluorescent 2 , 7 -dichlorofluorescein (DCF). Fluorescent signals were detected by the FACSCanto™ flow cytometer (BD). A total of 10,000 cells were analyzed per sample.

Apoptosis and Cell Proliferation Assays
To perform Annexin V and propidium iodide (PI) staining, irradiated MCF-7 cells and mammospheres were trypsinized, washed with 1× PBS, centrifuged and stained with the Annexin V-FITC antibody (20 min, room temperature) and PI (0.02 mg/mL; Sigma, P4170). The percentage of apoptotic cells was evaluated using the FACSCanto™ flow cytometer (BD). A total of 1 × 10 4 cells were recorded per condition in three independent experiments. In the cell proliferation assay, irradiated MCF-7 cells and mammospheres were trypsinized, washed with 1× PBS, centrifuged and stained with Ki67. The stained cells were analyzed using the FACSCanto™ flow cytometer (BD).

In Vivo Tumorigenicity Assay
Female SCID mice (6 to 8 weeks old; n = 5 per group) were maintained, according to the procedures and guidelines of the Institutional Animal Ethics Committee (NCCS). A total of 2 × 10 6 MCF-7 cells (wild type/shNrf2) were injected subcutaneously into the mammary fat pads of female SCID mice along with a 1:2 ratio of growth-factor-reduced Matrigel (BD Biosciences). These mice were also injected with β-estradiol (Sigma-Aldrich, St. Louis, MO, USA) and observed for 2 months for the development of breast tumors. All the mice were euthanized, according to the institute's ethical procedures, and tumors were collected for further analysis. The length and width of the tumors were measured using a vernier caliper and volumes were calculated using the following formula: Tumor volume = 1/2 (length × width 2 ). conversion, according to the manufacturer's protocol. Primers spanning two promoters of the Keap1 gene were designed using Methyl Primer Express (Thermo Scientific). Promoter 1 Forward: 5 -GAGTTTTGGYGGGGAATT-3 ; Reverse: 5 -CCCTACCRCCTAAAACCAA-3 . Bisulfite-modified DNA (100 ng) was amplified in a PCR mix containing 0.4 µM of forward and reverse primer, HotStarTaq Master Mix Kit (Qiagen, Germany: 203445). Methylation status analysis was performed by Quantification Tool for Methylation Analysis (QUMA) software.

Statistical Analysis
One-way analysis of variance (ANOVA), followed by Tukey's post hoc multiple comparisons tests by Prism software (GraphPad, San Diego, CA, USA), was used to analyze statistical significance. All the data values are presented as mean ± SE, reflecting the minimum of three independent determinations. Statistical significance was determined by comparing the treatments with untreated controls, and the significant differences are indicated as * p < 0.05, ** p < 0.01 and *** p < 0.001.

Fractionated Doses of Radiation Selectively Increase E-BCSC Population While Decreasing M-BCSC Population
Recent studies indicate that BCSCs exist in two phenotypes, i.e., epithelial (E-BCSC) and mesenchymal (M-BCSC), and BCSC plasticity plays a crucial role in future strategies for therapeutic resistance [19]. E-BCSCs characterized as ALDH + population are proliferative, locate in the tumor's hypoxic region and show the MET phenotype. On the other hand, M-BCSCs that express the CD44 + /24 − phenotype are primarily quiescent, located on the invasive front and have the EMT phenotype. Previous studies have shown an increase in CD44 + /24 − cells and high ALDH + characteristics of tumor-initiating or cancer stem cells in breast tumors and established cell lines after irradiation [20][21][22][23]. In our study, fractionated irradiation with 2 Gy x 3 days of γ-rays increased the population of ALDH + cells ( Figure 1A) but decreased CD44 + /24 − cells ( Figure 1B) in MCF-7 and MDA-MB-231 cells and their corresponding mammospheres. Since mammospheres render an enriched BCSC population [24], we characterized these mammospheres by quantifying embryonic stem cell markers, SOX2 and NANOG. Compared to the MCF-7 cells, MCF-7-derived mammospheres express significantly high levels of SOX2 and NANOG, indicating the enriched BCSC population ( Figure S1). Similar to ALDH activity, the expression of embryonic stem cell markers, i.e., SOX2 and NANOG in MCF-7 cells and mammospheres, and mammosphere formation efficiency (MFE) in MCF-7 cells was also increased upon exposure to fractionated doses of radiation ( Figure 1C,D). Collectively, these results suggest that exposure of fractionated doses to radiation induces the E-BCSC phenotype in mammospheres, which may contribute to radioresistance and promote tumor recurrence.

Fractionated Doses of Radiation Induce Cellular Plasticity by Regulating EMT
An increase in E-BCSC signature in our study prompted us to further analyze the EMT markers. Fractionated irradiation caused the induction of MET, as levels of the epithelial marker E-cadherin were observed to be increased and the levels of mesenchymal markers Vimentin, SLUG and SNAIL were found to be decreased significantly only in BCSCenriched mammospheres but not in MCF-7 cells (Figure 2A,B), thus inducing plasticity toward epithelial phenotype. GAPDH is used as loading control. All values are given as the mean ± SE, * p < 0.05, ** p < 0.01, *** p < 0.001, ****p <0.0001 vs. control. All images are representative of three independent experiments.

BCSCs with High ALDH + Activity Display Radioresistance Upon Exposure to Fractionated Irradiation
Although controversial, previous findings suggest that BCSCs might be less sensitive to irradiation than cancer cells in in vitro assays [4,5]. We used a clonogenic cell survival

BCSCs with High ALDH + Activity Display Radioresistance upon Exposure to Fractionated Irradiation
Although controversial, previous findings suggest that BCSCs might be less sensitive to irradiation than cancer cells in in vitro assays [4,5]. We used a clonogenic cell survival assay to analyze the relative radioresistance of BCSCs. A single-cell suspension of MCF-7 cells was plated and irradiated with an acute dose (6 Gy) and fractionated doses (2 Gy × 3 days) of γ-rays. Our clonogenic survival assay demonstrated significantly higher radioresistance in MCF-7 and MDA-MB-231 cells and their corresponding mammospheres upon exposure to fractionated doses of radiation compared to controls ( Figure 3A,B). Not only did the number of the colonies formed increase significantly after fractionated irradiation but also proliferative capacity, as indicated by Ki67 staining, was higher in these cells ( Figure 3C). Ionizing radiation significantly increased the proportion of these CSCs and also showed enhanced proliferation shortly after treatment, further resulting in rapid tumor repopulation [25]. As there was an increase in the proliferation in cancer cells and mammospheres after fractionated irradiation, we further assessed apoptosis and the expression of antiand proapoptotic genes, BCL2 and BAX. Although there was no significant change in the Annexin V + apoptotic population in MCF-7 cells and mammospheres after fractionated irradiation compared to their respective controls ( Figure 3D), a significant increase in the BCL2/BAX ratio was observed at the protein levels, further supporting radioresistance in these cells ( Figure 3E).
Cells 2021, 10, x FOR PEER REVIEW 9 of 23 change in the Annexin V + apoptotic population in MCF-7 cells and mammospheres after fractionated irradiation compared to their respective controls ( Figure 3D), a significant increase in the BCL2/BAX ratio was observed at the protein levels, further supporting radioresistance in these cells ( Figure 3E).

The Emergence of Radioresistance Is Associated with High Migratory Potential and Tumorigenicity in Cancer Cells
To analyze whether breast cancer cells irradiated with fractionated doses of radiation have functional characteristics of BCSCs, we examined their cell migration potential in vitro and tumorigenic properties in vivo. Compared to the controls and an acute dose, a

The Emergence of Radioresistance Is Associated with High Migratory Potential and Tumorigenicity in Cancer Cells
To analyze whether breast cancer cells irradiated with fractionated doses of radiation have functional characteristics of BCSCs, we examined their cell migration potential in vitro and tumorigenic properties in vivo. Compared to the controls and an acute dose, a significant increase in migration efficiency was observed in cells irradiated with the fractionated doses of radiation in the scratch wound assay ( Figure 4A). Further, tumors in mice derived from MCF-7 cells irradiated with fractionated doses of radiation weighed significantly more than tumors derived from nonirradiated or acute-dose-irradiated MCF-7 cells ( Figure 4B,C). Consistent with the in vitro results, analysis of the xenograft tumors derived from tumor cells irradiated with fractionated doses also showed enhanced ALDH activity ( Figure 4D). Overall, these data demonstrate that fractionated dose exposure enhances migration potential in vitro and increases tumorigenicity by elevating the ALDH + population in vivo. mice derived from MCF-7 cells irradiated with fractionated doses of radiation weighed significantly more than tumors derived from nonirradiated or acute-dose-irradiated MCF-7 cells (Figure 4B,C). Consistent with the in vitro results, analysis of the xenograft tumors derived from tumor cells irradiated with fractionated doses also showed enhanced ALDH activity ( Figure 4D). Overall, these data demonstrate that fractionated dose exposure enhances migration potential in vitro and increases tumorigenicity by elevating the ALDH + population in vivo.  Diehn et al. [5] showed that CSCs in breast tumors contain low ROS levels and enhanced ROS defenses compared to their nontumorigenic progeny, and these differences appear to be critical for maintaining stem cell function, which could contribute to tumor radioresistance. Previous studies have shown the involvement of Nrf2 in chemoresistance in BCSCs [12,13], hence we hypothesized that Nrf2 could also play a significant role in the radioresistance of BCSCs. We first determined the levels of ROS in MCF-7 cells and mammospheres irradiated with fractionated doses of radiation. We did not see any change in the ROS levels in these cells compared to their respective controls. However, an acute dose of radiation increased the levels of ROS in MCF-7 cells as well as in mammospheres (Figure 5A). Western blot and qRT-PCR analysis revealed that Nrf2 expression ( Figure 5B,C), activity ( Figure 5D), as well as its targets HO1 and NQO1 ( Figure 5E

Keap1-Nrf2 and not Bach1-Nrf2 Signaling Plays a Role in the Maintenance of Radioresistant ALDH + BCSCs
Diehn et al. [5] showed that CSCs in breast tumors contain low ROS levels and enhanced ROS defenses compared to their nontumorigenic progeny, and these differences appear to be critical for maintaining stem cell function, which could contribute to tumor radioresistance. Previous studies have shown the involvement of Nrf2 in chemoresistance in BCSCs [12,13], hence we hypothesized that Nrf2 could also play a significant role in the radioresistance of BCSCs. We first determined the levels of ROS in MCF-7 cells and mammospheres irradiated with fractionated doses of radiation. We did not see any change in the ROS levels in these cells compared to their respective controls. However, an acute dose of radiation increased the levels of ROS in MCF-7 cells as well as in mammospheres ( Figure 5A). Western blot and qRT-PCR analysis revealed that Nrf2 expression ( Figure 5B,C), activity ( Figure 5D), as well as its targets HO1 and NQO1 ( Figure 5E,F), increased significantly when treated with fractionated doses of radiation. We observed a significant decrease in the expression of Keap1, and there was no change in the expression of Bach1, MCF-7 and MDA-MB-231 cells and their corresponding mammospheres irradiated with fractionated doses ( Figure 5G,H), indicating that Keap1-mediated Nrf2 degradation is impaired, leading to the stabilization of Nrf2 and its nuclear accumulation [10,12]. The reduced level of ROS in our study could therefore be attributed to the activation of the antioxidant defense mechanism. fractionated doses ( Figure 5G,H), indicating that Keap1-mediated Nrf2 degradation is impaired, leading to the stabilization of Nrf2 and its nuclear accumulation [10,12]. The reduced level of ROS in our study could therefore be attributed to the activation of the antioxidant defense mechanism.

Inhibition of Nrf2 Concealed Radioresistance, Tumorigenesis and Induced Apoptosis Via Reducing BCSC Population
To further investigate the role of Nrf2 in radioresistance, Nrf2 was knocked down in MCF-7 cells (shNrf2). These cells showed a 55% reduction in Nrf2 transcripts levels (Figure S2). A 50% reduction in the population of ALDH + cells was observed in Nrf2-knockdown mammospheres and MCF-7 cells after fractionated irradiation ( Figure 6A). As a phenotypic effect, stable silencing of Nrf2 also resulted in the inhibition of mammosphere formation efficiency by two-fold in MCF-7 cells ( Figure 6B). A reduction in the levels of SOX2, KLF4 and NANOG in these knockdown cells after irradiation indicated the role of Nrf2 in the suppression of BCSC population ( Figure 6C). Tumorigenicity in SCID mice

Inhibition of Nrf2 Concealed Radioresistance, Tumorigenesis and Induced Apoptosis via Reducing BCSC Population
To further investigate the role of Nrf2 in radioresistance, Nrf2 was knocked down in MCF-7 cells (shNrf2). These cells showed a 55% reduction in Nrf2 transcripts levels ( Figure S2). A 50% reduction in the population of ALDH + cells was observed in Nrf2knockdown mammospheres and MCF-7 cells after fractionated irradiation ( Figure 6A). As a phenotypic effect, stable silencing of Nrf2 also resulted in the inhibition of mammosphere formation efficiency by two-fold in MCF-7 cells ( Figure 6B). A reduction in the levels of SOX2, KLF4 and NANOG in these knockdown cells after irradiation indicated the role of Nrf2 in the suppression of BCSC population ( Figure 6C). Tumorigenicity in SCID mice was decreased after injection of the irradiated Nrf2 knockdown cells. A significant decrease in tumor size ( Figure 6D) as well as the percentage of ALDH + population was observed in these tumors compared to the corresponding control ( Figure 6E). Further, a reduction in clonogenicity ( Figure 6F) and a significantly higher number of Annexin-V-/PI-positive cells were observed compared to their respective controls in shNrf2 mammospheres and MCF-7 cells irradiated with fractionated doses ( Figure 6G). Thus, these results suggest that Nrf2 plays a crucial role in the acquisition of radiation resistance in BCSCs. was decreased after injection of the irradiated Nrf2 knockdown cells. A significant decrease in tumor size ( Figure 6D) as well as the percentage of ALDH + population was observed in these tumors compared to the corresponding control ( Figure 6E). Further, a reduction in clonogenicity ( Figure 6F) and a significantly higher number of Annexin-V-/PIpositive cells were observed compared to their respective controls in shNrf2 mammospheres and MCF-7 cells irradiated with fractionated doses ( Figure 6G). Thus, these results suggest that Nrf2 plays a crucial role in the acquisition of radiation resistance in BCSCs.

miR200a and not Promoter Methylation of Keap1 is Involved in Radioresistance of BCSC
Since we observed a significant decrease in the expression of Keap1 at mRNA and protein levels, we further investigated its regulation at the epigenetic level, especially the methylation status of the Keap1 promoter by bisulfite sequencing [26]. We did not observe any change in the methylation status of the CpGs region in the Keap1 promoter, indicating that Keap1 promoter methylation may not be the key event in Nrf2 stabilization ( Figure  7A,B). We next examined the role of the miR-200 family as it targets a conserved region in the Keap1 3′-UTR [27]. We observed no change in the expression of miR-141 but a significant increase in the expression of miR-200a, 1.4-fold in mammospheres and 1.85-fold in MCF-7 cells irradiated with the fractionated dose of radiation by RT-PCR ( Figure 7C,D). Collectively, these results indicate that Keap1 downregulation could be due to increased miR200a; however, more studies are required to confirm the role of miR200a in this context.

miR200a and not Promoter Methylation of Keap1 is Involved in Radioresistance of BCSC
Since we observed a significant decrease in the expression of Keap1 at mRNA and protein levels, we further investigated its regulation at the epigenetic level, especially the methylation status of the Keap1 promoter by bisulfite sequencing [26]. We did not observe any change in the methylation status of the CpGs region in the Keap1 promoter, indicating that Keap1 promoter methylation may not be the key event in Nrf2 stabilization ( Figure 7A,B). We next examined the role of the miR-200 family as it targets a conserved region in the Keap1 3 -UTR [27]. We observed no change in the expression of miR-141 but a significant increase in the expression of miR-200a, 1.4-fold in mammospheres and 1.85-fold in MCF-7 cells irradiated with the fractionated dose of radiation by RT-PCR ( Figure 7C,D). Collectively, these results indicate that Keap1 downregulation could be due to increased miR200a; however, more studies are required to confirm the role of miR200a in this context.

Discussion
Radiation can induce cancer cell death by generating ROS and DNA damage; however, it is inefficient in targeting CSCs, which are largely responsible for therapy resistance, tumorigenesis and tumor recurrence [28][29][30]. Our study demonstrates that fractionated doses of radiation enhanced the E-BCSC marker ALDH + and transcription factors of embryonic stem cells in BCSC-enriched mammospheres, indicating the E-BCSC phenotype, which is proliferative in nature. BCSC plasticity plays a crucial role in therapy resistance. BCSCs exhibit plasticity, which transitions between quiescent mesenchymal-(M-BCSCs) and proliferative epithelial-like (E-BCSCs) states [31]. An increase in E-BCSCs such as ALDH+ population and E-cadherin, indicative of MET, and a decrease in M-BCSCs such as CD44 + /24 − population, the mesenchymal markers Vimentin, SNAIL and SLUG, demonstrated that fractionated doses of radiation increase the epithelial type of BCSCs [24,31,32]. Thus, these results support the notion that BCSC markers are not restricted to a particular population but change according to their plasticity based on the therapy. Hence, plasticity from M-BCSCs to E-BCSCs contributes to radioresistance. Since NANOG, SOX2 and KLF4 are essential for converting tumor cells into aggressive stem-

Discussion
Radiation can induce cancer cell death by generating ROS and DNA damage; however, it is inefficient in targeting CSCs, which are largely responsible for therapy resistance, tumorigenesis and tumor recurrence [28][29][30]. Our study demonstrates that fractionated doses of radiation enhanced the E-BCSC marker ALDH + and transcription factors of embryonic stem cells in BCSC-enriched mammospheres, indicating the E-BCSC phenotype, which is proliferative in nature. BCSC plasticity plays a crucial role in therapy resistance. BCSCs exhibit plasticity, which transitions between quiescent mesenchymal-(M-BCSCs) and proliferative epithelial-like (E-BCSCs) states [31]. An increase in E-BCSCs such as ALDH+ population and E-cadherin, indicative of MET, and a decrease in M-BCSCs such as CD44 + /24 − population, the mesenchymal markers Vimentin, SNAIL and SLUG, demonstrated that fractionated doses of radiation increase the epithelial type of BCSCs [24,31,32]. Thus, these results support the notion that BCSC markers are not restricted to a particular population but change according to their plasticity based on the therapy. Hence, plasticity from M-BCSCs to E-BCSCs contributes to radioresistance. Since NANOG, SOX2 and KLF4 are essential for converting tumor cells into aggressive stem-like cells, an increase in the expression of these markers in our study after irradiation further supports the increased cancer stem cell population.
Emerging evidence indicates that Nrf2 plays a crucial role in CSC survival and resistance [33]. It is shown to be involved in chemotherapeutic drug resistance due to enhanced antioxidant capacity and detoxification of anticancer agents [14,34,35]. However, the involvement of the Nrf2-Keap1 axis in radioresistance of BCSCs is poorly understood. A strong association between low levels of ROS and enhanced antioxidant defense in BCSC radioresistance reported by Diehn et al. [5] prompted us to further investigate the role of Nrf2. Enhanced expression of Nrf2 and its downstream genes HO1 and NQO1 after irradiation in breast cancer cells and their corresponding mammospheres ascertains the involvement of Nrf2 in radioresistance. A recent report has shown that Nrf2 enhances ALDH + E-BCSCs [24]. This supports our results, as we have observed a decrease in the ALDH + E-BCSCs after Nrf2 inhibition. A decrease in embryonic stem cell markers, colony and sphere formation ability and reduced tumorigenicity after Nrf2 knockdown further indicate that Nrf2 is involved in the reprogramming process, and Nrf2 signaling is an important target for radiation resistance of BCSCs.
In the current study, Nrf2 appears to be regulated by Keap1 as we observed a decrease in the Keap1 levels with no change in the expression of either GSK-3β ( Figure S3) [36] or Bach1. Additionally, as Bach1 binds to HO1 [10,11], an increase in the levels of HO1 in our study further confirms that Bach1 does not play a role in the regulation of Nrf2. Loss of Keap1 function is shown to mediate Nrf2 stabilization and is often associated with reduced drug sensitivity in several cancers [37][38][39]. A reduction in Keap1 expression with a concomitant increase in the expression of Nrf2 and its downstream targets HO1 and NQO1 clearly demonstrates the role played by Keap1 in Nrf2 regulation in the facilitation of acquired radioresistance. Hence, we tried to understand the mechanism of Keap1 regulation in this study.
Besides mutations through cysteine residues, epigenetic mechanisms, particularly the promoter hypermethylation [26], and miRNAs are the main regulators of Keap1. We did not see any change in the promoter methylation status of Keap1 after irradiation, which suggested that irradiation may regulate Keap1 post-transcriptionally rather than epigenetically. Hence, we further studied the role of the miR200 family as it is known to be involved in the regulation of Keap1. A significant increase in the transcript levels of miR200a indicates its role in the regulation of Keap1 in the radioresistance of BCSCs. Furthermore, reports from other studies have shown that miR200a suppresses the expression of transcriptional factors ZEB1/2 and inhibits the transition from the epithelial-to-mesenchymal phenotype [40]. This further strengthens and supports our studies where miR200a could be responsible for the inhibition of Keap1 as well as EMT in BCSC-enriched mammospheres.
In conclusion, the current study provides interesting insights into the mechanism by which fractionated doses of radiation increases radioresistance in the BCSC population.
Our results indicate the enrichment of the E-BCSC phenotype. The regulation of Nrf2 in irradiated conditions occurs via the downregulation of Keap1 and not by GSK3β or Bach1. We provide mechanistic insight into the regulation of Keap1, possibly via posttranscriptional modification through miR200a and not via promoter methylation. Although the current study is limited to only the higher expression of miR200a, and given its potential for therapeutic purposes, additional mechanistic studies regarding its role in Keap1 inhibition and thus radioresistance is highly warranted. Nevertheless, alteration in the Nrf2-Keap1 pathway establishes relationships between radioresistance and BCSCs.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.